Merge branch 'main' into py38-compat

.github/workflows/docs.yml

@@ -14,7 +14,6 @@ jobs:
     uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
     with:
       repository: pytorch/vision
-      upload-artifact: docs
       script: |
         set -euo pipefail
 
@@ -40,42 +39,9 @@ jobs:
         pip install --progress-bar=off -r requirements.txt
         echo '::endgroup::'
         
-        echo '::group::Build HTML docs'
-        # The runner does not have sufficient memory to run with as many processes as their are
+        # The runner does not have sufficient memory to run with as many processes as there are
         # cores (`-j auto`). Thus, we limit to a single process (`-j 1`) here.
         sed -i -e 's/-j auto/-j 1/' Makefile
         make html
-        echo '::endgroup::'
         
-        mv build/html "${RUNNER_ARTIFACT_DIR}"
-
-  upload-preview:
-    if: github.event_name == 'pull_request'
-    needs: [build]
-    runs-on: [self-hosted, linux.2xlarge]
-    steps:
-      - uses: actions/download-artifact@v3
-        with:
-          name: docs
-
-      - name: Upload docs preview
-        uses: seemethere/upload-artifact-s3@v5
-        with:
-          retention-days: 14
-          s3-bucket: doc-previews
-          if-no-files-found: error
-          path: html
-          s3-prefix: pytorch/vision/${{ github.event.pull_request.number }}
-
-      # The upload below duplicates the upload from above, but to a different path. This is needed since we are in the
-      # process of changing the path, but want to keep the disruption to a minimum.
-      # See https://github.com/pytorch/test-infra/issues/3894
-      # After a grace period, we can delete this again
-      - name: Upload docs preview
-        uses: seemethere/upload-artifact-s3@v5
-        with:
-          retention-days: 14
-          s3-bucket: doc-previews
-          if-no-files-found: error
-          path: html
-          s3-prefix: pytorch/pytorch/vision/${{ github.event.pull_request.number }}
+        mv build/html/* "${RUNNER_DOCS_DIR}"

.github/workflows/test-windows.yml

@@ -30,6 +30,8 @@ jobs:
     with:
       repository: pytorch/vision
       runner: ${{ matrix.runner }}
+      gpu-arch-type: ${{ matrix.gpu-arch-type }}
+      gpu-arch-version: ${{ matrix.gpu-arch-version }}
       timeout: 120
       script: |
         set -euxo pipefail
@@ -38,14 +40,4 @@ jobs:
         export GPU_ARCH_TYPE=${{ matrix.gpu-arch-type }}
         export GPU_ARCH_VERSION=${{ matrix.gpu-arch-version }}
         
-        # TODO: Port this to pytorch/test-infra/.github/workflows/windows_job.yml
-        export PATH="/c/Jenkins/Miniconda3/Scripts:${PATH}"
-        
-        if [[ $GPU_ARCH_TYPE == 'cuda' ]]; then
-          # TODO: This should be handled by the generic Windows job the same as its done by the generic Linux job
-          export CUDA_HOME="/c/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v${{ matrix.gpu-arch-version }}"
-          export CUDA_PATH="${CUDA_HOME}"
-          export PATH="${CUDA_PATH}/bin:${PATH}"
-        fi
-        
         ./.github/scripts/unittest.sh

test/test_transforms_v2.py

@@ -2137,3 +2137,38 @@ def test_no_warnings_v1_namespace():
         from torchvision.datasets import ImageNet
     """
     assert_run_python_script(textwrap.dedent(source))
+
+
+class TestLambda:
+    inputs = pytest.mark.parametrize("input", [object(), torch.empty(()), np.empty(()), "string", 1, 0.0])
+
+    @inputs
+    def test_default(self, input):
+        was_applied = False
+
+        def was_applied_fn(input):
+            nonlocal was_applied
+            was_applied = True
+            return input
+
+        transform = transforms.Lambda(was_applied_fn)
+
+        transform(input)
+
+        assert was_applied
+
+    @inputs
+    def test_with_types(self, input):
+        was_applied = False
+
+        def was_applied_fn(input):
+            nonlocal was_applied
+            was_applied = True
+            return input
+
+        types = (torch.Tensor, np.ndarray)
+        transform = transforms.Lambda(was_applied_fn, *types)
+
+        transform(input)
+
+        assert was_applied is isinstance(input, types)

test/transforms_v2_kernel_infos.py

@@ -321,6 +321,9 @@ def reference_resize_bounding_box(bounding_box, *, spatial_size, size, max_size=
     old_height, old_width = spatial_size
     new_height, new_width = F._geometry._compute_resized_output_size(spatial_size, size=size, max_size=max_size)
 
+    if (old_height, old_width) == (new_height, new_width):
+        return bounding_box, (old_height, old_width)
+
     affine_matrix = np.array(
         [
             [new_width / old_width, 0, 0],

torchvision/transforms/v2/_misc.py

@@ -32,10 +32,12 @@ class Lambda(Transform):
         lambd (function): Lambda/function to be used for transform.
     """
 
+    _transformed_types = (object,)
+
     def __init__(self, lambd: Callable[[Any], Any], *types: Type):
         super().__init__()
         self.lambd = lambd
-        self.types = types or (object,)
+        self.types = types or self._transformed_types
 
     def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
         if isinstance(inpt, self.types):

torchvision/transforms/v2/functional/_geometry.py

@@ -179,7 +179,9 @@ def resize_image_tensor(
     num_channels, old_height, old_width = shape[-3:]
     new_height, new_width = _compute_resized_output_size((old_height, old_width), size=size, max_size=max_size)
 
-    if image.numel() > 0:
+    if (new_height, new_width) == (old_height, old_width):
+        return image
+    elif image.numel() > 0:
         image = image.reshape(-1, num_channels, old_height, old_width)
 
         dtype = image.dtype
@@ -210,9 +212,19 @@ def resize_image_pil(
     interpolation: Union[InterpolationMode, int] = InterpolationMode.BILINEAR,
     max_size: Optional[int] = None,
 ) -> PIL.Image.Image:
+    old_height, old_width = image.height, image.width
+    new_height, new_width = _compute_resized_output_size(
+        (old_height, old_width),
+        size=size,  # type: ignore[arg-type]
+        max_size=max_size,
+    )
+
     interpolation = _check_interpolation(interpolation)
-    size = _compute_resized_output_size(image.size[::-1], size=size, max_size=max_size)  # type: ignore[arg-type]
-    return _FP.resize(image, size, interpolation=pil_modes_mapping[interpolation])
+
+    if (new_height, new_width) == (old_height, old_width):
+        return image
+
+    return image.resize((new_width, new_height), resample=pil_modes_mapping[interpolation])
 
 
 def resize_mask(mask: torch.Tensor, size: List[int], max_size: Optional[int] = None) -> torch.Tensor:
@@ -235,6 +247,10 @@ def resize_bounding_box(
 ) -> Tuple[torch.Tensor, Tuple[int, int]]:
     old_height, old_width = spatial_size
     new_height, new_width = _compute_resized_output_size(spatial_size, size=size, max_size=max_size)
+
+    if (new_height, new_width) == (old_height, old_width):
+        return bounding_box, spatial_size
+
     w_ratio = new_width / old_width
     h_ratio = new_height / old_height
     ratios = torch.tensor([w_ratio, h_ratio, w_ratio, h_ratio], device=bounding_box.device)